from February 20, 2003 2:00-3:00 PM Eastern
Copyright © by
International Society for Complexity, Information, and Design 2003.
Our guest speaker today is biochemist James
Shapiro. Dr. Shapiro
is Professor in the Department of Biochemistry and Molecular
the University of Chicago. He received a B.A. in English Literature
from Harvard University in 1964 and a Ph.D. in Genetics from Cambridge
University in 1968. Dr. Shapiro's most recent research has been centered
around the concept of "natural genetic engineering" which
is highlighted in his thought provoking paper A
21st Century View of evolution
Dr. Shapiro asked that I post this opening statement.
It is a summary of his ideas...
Genomes constitute the long-term information storage organelles of
every cell and are hierarchically organized as systems assembled
from DNA modules. Each genome has a computational system architecture
formatted by sequence elements that do not code for proteins. Whole-genome
sequencing indicates that rearrangement of genetic modules plus duplication
and reuse of existing genomic systems are fundamental events in evolution.
Molecular discoveries about mechanisms of DNA restructuring show
that cells possess the Natural Genetic Engineering functions necessary
for evolutionary change by rearranging genomic components. These
natural genetic engineering functions are sensitive to biological
inputs, and their non-random operations help explain how novel system
architectures can arise in evolution. Darwin himself recognized the
potential for major hereditary variation "as leading to permanent
modifications of structure independently of natural selection" (Origin
of Species, 6th edition, Chapter XV, p. 395). By integrating the
emerging concepts of cellular computation and decision-making, genome
system architecture, and natural genetic engineering, we are
now in a position to frame novel 21st Century evolutionary theories
in terms of systems engineering.
I am now going to hand the talk over to Dr. Shapiro. Participants can
start sending in questions.
James, we met the first time at Wheaton College at a symposium in the
spring of 1997 featuring principally you and Michael
Behe along with
Paul Nelson and David Hull. At the time I asked you about the origin
of such "natural genetic engineering" systems. As I recall,
you indicated that this was not really the problem you were addressing.
Have you thought any more about this problem? Specifically, how do
such systems arise that can take over their evolution? And how much
complexity do they require? Are you confident that non-teleological
mechanisms can account for the rise of natural genetic engineering
systems, and if so why?
I am not sure how to answer your question. All existing living organisms
possess natural genetic engineering capabilities. So they must be
pretty fundamental. Any self-organizing evolving system has to have
the capacity to alter its information store. That's what they do.
Where they come from in the first place is not a question we can
realistically answer now, any more than we can explain the origin
of the first cells.
Dr Shapiro, Natural genetic engineering does change our views on evolution.
Is there any evidence that genetic circuits able to sense the environment
and change the genome accordingly exist in higher metazoans? Would
they be able to give an organism heritable changes? And finally,
is there anyone doing serious work on this issue in particular?
There is a lot of work on this issue. We know that cells activate natural
genetic engineering functions in response to various inputs, particularly
stresses (what McClintock called "genome shock"). In certainly
highly evolved situations, like the immune system, the responsiveness
of these systems is quite extraordinary.
I'm curious about "their non-random operations." Does 'non-random'
suggest that the very instructions for all possible morphological changes
are front loaded or pre-programmed into living things, needing only
a given catalyst to get things going?
No. Non-random means that they operate under certain conditions (e.g.
after genome damage or viral infection) and that these systems make
characteristic kinds of changes. When a retrovirus-like element inserts
in a new genomic location, it carries with it a defined set of regulatory
signals that can affect the reading of nearby DNA sequences in very
particular ways. This is an example of non-randomness. In addition,
some changes (such as those in the immune system) can be targeted
to specific locations by the presence of particular signals in the
DNA or by activation of transcription. These phenomena show us that
cells are capable of altering their genomes in non-random but not
rigidly specified or pre-determined ways.
Dr. Shapiro, are there other researchers/labs that currently exemplify
your concept of a 21st century view of evolution? What type of work
are they doing and could you briefly describe some of the interesting
results that you've come across in your research?
Goodness. If you look at my web site (http://shapiro.bsd.uchicago.edu),
there are several papers available that give lists of references.
There are books on the roles of transposable elements in evolution.
Springer just published a book on "Evolution as Computation" (Landweber
and Winfree, eds), and the NY Academy of Sciences has published a
couple of volumes where these ideas have been explored at length
(L. Caporale, ed. 1999, "Molecular Strategies in Biological
Evolution: and one just published on Epigenesis at the end of 2002).
Does this help?
Could Natural genetic engineering cause an organism that is exposed
to a drastic environmental change (like when a cataclismic event
ocurs on earth) rearange the genetic code to adapt to the new environment?
Has their been any examples of natural genetic engineering causing
changes on the macro level?
In addition, in answer to Micah's question, I shouild add that recent
papers on genome sequences incoporate and exemplify these ideas.
The recent Nature paper comparing the mouse and human genomes is
a good example.
In response to Yaakov, we know of major genome rearrangements in response
to various kinds of crisis. McClintock documented this in her maize
plants subject to repeated cycles of chromosome breakage. When plant
cells are put in tissue culture and plants are then regenerated,
major genome restructurings are often observed. When matings occur
between different Drospophila populations, widespread chromosome
rearrangements, mutations and spread of transposable elements can
occur. Some of these systems studied in the laboratory correspond
to differences between species in nature. There are lots of similarities
between the mouse and human genomes, for example, but we know that
literally tens of thousands of retrotransposition events separate
the two species.
I understand the immune system adaptation, but my question was different.
I was asking about heritable changes offered by natural genetic engineering.
For instance, If levels of sea water rise on the planet and we are
essentially left with a "waterworld". Would we expect genomic
rearrangements to happen in the gametes of terrestrial animals so
that their offsprings are essentially born with respiratory systems
closer to those of marine animals?
Waterworld scenario implausible to start (not enough available water
to do that) ... and we already have clues to one such transition,
whales, which involved many millions of years again. A thin slice
of time would show only modest microevolutionary wiggling in the
genome. Would Dr. Shapiro care to comment?
What I can say is that organisms have diversified to occupy virtually
all available ecological niches. When we look at their genomes, we
see similar systems encoded in the DNA but often rearranged and adapted
for specific purposes in ways that we know can occur in the laboratory.
So it seems reasonable to conclude that these natural genetic engineering
processes have participated in the adaptations to new environments.
In response to Downard, I think the main point is that we often see
major scrambling of genomes between relatively closely related organisms
(e.g. mosquitos). Moreover, these rearrangements are abrupt, cut-and-splice
types of changes, not the accumulation of many small modifications.
So the genome comparisons mirror what is observed experimentally,
namely that cells can rapidly reorganize their genomes.
Dr. Shapiro, How do you conceptualize information? Is it real or only
abstract? Could it have structure, which unfolds as an active agent
in the structuring of physical processes?
This question is too philosophical for me. There are many kinds of
information in cells, and only some of them are stored in the DNA.
Perhapos you can phrase the question more specifically.
When the major genome restructuring occur in maize plants occur, are
they random restructuring or non-random. Additionally, are the chromosome
rearangements, mutations and spread of transosable elements observed
in drosophila random or non-random producing more advanced offspring?
The changes occur non-randomly in the sense that they follow certain
predilections (e.g. some mobile elements insert near the start sites
of transcription, others prefer to insert in protein coding sequences).
Often these changes have major effects on phenotype. If we set up
the situation properly, we can often see quite high frequencies of
changes that are advantageous to the organism, as in my own work
on adaptive mutation in bacteria. Most of this experimental work
has been done with microbes, and there we know for certain that important
adaptive traits (e.g. antibiotic resistance) have evolved by natural
genetic engineering processes.
Dr. Shapiro, you mentioned that there are several types of information
stored in cells and only some of it in DNA. Could you explain to
some of us non-biologists what some of the other sources of information
are within the cell?
For example, cells have information about their physiological status.
In E. coli, the availability of glucose is indicated by the intracellular
level of cyclic AMP, a small cytiplasmic molecule. This information
can influence both the expression and alteration of genomic information,
but it is different in nature. Dos that help?
Re the whale example, the recent paleontology indicates many millions
of years for just salt adaptation to accumulate, as indicated by
the habitat suggested by their taphonomy (various papers in Nature
on these). A paleontological perspective adds a further dimension
to discussions of genetic activity in the individual and population
level among extant organisms. Again, would Dr. Shapiro comment?
I don't know very much about whales. As I understand it, the paleontological
record is often very punctuated with long periods of stasis followed
by rapid periods of change. This is exactly what one would expect
if natural genetic engineering occurs episodically in response to
Recently, it has been shown that neurons of a single organism have
different numbers of chromosomes. What happens is that different
neurons kick out chromosomes ending up with a population of cells
with a heterogenous genetic background. Could this be a manifestation
of natural genetic engineering you are studying?
Probably. Although it is far from a general mechanism, very predictable
changes to the genome happen during the developmenbt of many multicellular
Dr. Shapiro, normally, one cannot take a section of computer code,
remove it, and drop it in somewhere else without causing a loss,
or decrease, of functionality for at least part of the application.
I would expect to see a good deal of this sort of problem in nature.
If I were to observe constructive changes as a result of what appears
to be a scrambled genome (or computer code), I would conclude that
the code was built to handle certain aspects of scrambling, but not
others. In other words, the regulatory regions of the genome would
be expected to already be engineered to handle scrambling within
the built-in specs. It seems to me there is a way to test this, by
performing our own scrambling of genomes in the lab and then observing
how viable the organism was. I'm not talking about merely dropping
a transposon onto a gene, but something more on the scale of what
we observe between mouse and humans. Is this being done? I personally
would expect genetic disaster, which would imply that the scrambling
may only scrambling
may only be properly executed under the supervision of ID.
In response to Kirk, I should direct you to a web page on self-evolving
computer programs at Stanford (associated with John Koza, I believe).
It turns out that shuffling of fragments of programs is actually
a very good strategy for developing powerful new programs and algorithms.
So the electronic systems seem to like the biological way of doing
things. The key point is that large, already functional assemblages
are allowed to recombine in novel ways.
As is known, any well-defined genus of bacteria, especially Gram-negative
bacteria, has more rapidly growing and less-rapidly growing species
- e.g. in Pseudomonas, it's Ps. fluorescence and Ps. aeruginosa.
A more general differentiation would be bacteria with explosive metabolism
and extremely slowly growing oligotrophs, such as Hyphomicrobium,
Prosthecomicrobium, etc. This type of differentiation may be explained
by ecological specifics of these microbial forms. In this view, do
you see any "universals" in re-arrangement of genetic modules
in whole-genome sequencing?
I don't know how to answer this question. Where we understand a particular
adaptation, as in bacterial virulence, we have lots of evidence for
natural genetic engineering in its evolution. Moreover, we see that
the genomes of different organisms have characteristic "system
architectures," in the same way as different types of computers
have distinct architectures. I hypothesize that large-scale adaptive
changes may be correlated with major systems in genomic system architecture,
as in the widespread change of repetitive DNA. But right now that
is only a hypothesis.
Metazoans are not able to carry out horizontal transfer of genetic
material to the extent that single-celled organisms do (or so it
is thought). Would this alter the extent and/or strategy of natural
genetic engineering in the two groups?
Naturally. But we may not fully appreciate the extent of lateral DNA
transfer among metazoans. After all, they are continually eating
each other and spilling their DNA all over the place, and animal
cells are wonderfully adept at taking up DNA and integrating it into
Dr. Schapiro. Do you expect that the 21st century evolutionary theory
you are working on will be able to explain speciation based on currently
In addition, certain authors argue that lateral DNA transfer has been
a major force in eukaryotic evolution, and I believe Mike Syvanen has
edited a book on this topic.
In response to Glenn, the answer is yes, in part. No one has created
true species by artifical selection, but we do accomplish that by
forced hybridizations and changing the ploidy of organisms. These
are conditions that inherently change the structure of the genome
and also activate natural genetic engineering functions.
I do think that there is an impressive amount of flexibility built
into any genome. I'm not familiar with John Koza's work, but I have
writen software for Defense Research (Canada). If sections of the
software were to be rearranged, there would have to be rules as to
where the cutting and splicing would be done. If the software itself
was in charge of this, the rules would have to be written in. The
only real way around this is to make functional sequence space, for
the genome, so large that one has a good chance of success. In other
words, I wonder of Koza does not utilize very low level functions,
unlike what we see in organic life.
I can't comment in detail on the software engineering systems. But
we can see how our genomes confront the issue of encoding infinity.
It occurs in the immune system, where a limitless variety of protein
coding sequences have to be generated from a finite DNA complement.
That is achieved by a highly integrated genetic engineering process
that includes combinatorial rearrangements, synthesis of new sequences,
and rather highly targeted improvement of the initial products so
that they work better over time. We can learn a lot about information
processing from our lymphocytes.
Well, that is it for today's chat. ISCID would like to thank Dr. Shapiro
for his thoughtful responses and his willingness to be with us today.
My pleasure. I only wish I were more practiced and could have givenmore
detailed answers. Please look at the web site.
Many thanks, Micah. I hope this was satisfactory and a least moderately
Yes it was....I can attest to that!
discussion with James Shapiro.
Copyright © by
International Society for Complexity, Information, and Design 2003.